Dolphin D: Improving Automation of Targeted Metabolomics in Multi-matrix Datasets of $$^$$H-NMR Spectra

Nuclear magnetic resonance (NMR) is one of the main tools applied in the field of metabolomics. Extracting all the valuable information from large datasets of \(^1\)H-NMR spectra is a huge challenge for high throughput metabolomics analysis. The tools that currently exist to improve signal assignment and metabolite quantification do not have the versatility of allowing the quantification of unknown signals or choosing different quantification approaches in the same analysis. Moreover, graphical features and informative outputs are needed in order to be aware of the reliability of the final results in a field where position shifting, baseline masking and signal overlap may produce errors between samples. Here we present a software package called Dolphin 1D, which aim is to improve targeted metabolite analysis in large datasets of \(^1\)H-NMR by combining user interactivity with automatic algorithms. Its performance has been tested on a multi-matrix set composed by total serum, urine, liver aqueous extracts and brain aqueous extracts of rat. Our strategy pretends to offer a useful solution for every kind of matrix, avoiding black-box processes and subjectivities user-user in automatic signal quantification.